Detailed flow, hydrometeor and lightning characteristics of an isolated, hail producing thunderstorm during COPS

Transcription

1 Detailed flow, hydrometeor and lightning characteristics of an isolated, hail producing thunderstorm during COPS, Martin Hagen, Hartmut Höller, Hans Volkert Institut für Physik der Atmosphäre, DLR Oberpfaffenhofen Folie 1

2 Results from previous studies of IOP 8b (15 July 2007) Observation Modelling Kottmeier et al. 2008, MetZet Aoshima et al. 2008, MetZet Kalthoff et al. 2009, AR Behrendt et al. 2011, QJRMS General observation results with special focus on convection initiation Evaluation of triggering mechanism (radar convergence line) Evaluation of moist condition Barthlott et al. 2009, AR Kirshbaum et al. 2010, JAS Richard et al. 2011, QJRMS Barthlott et al. 2011, QJRMS no deep convection by using COSMO-DE but triggering of convection matches well Intercomparison of different model results Comparison with observed moisture, changing of Meso-NH-model parameters current step: evaluation of mature state and dissipation > Folie 2 Folie 2

3 Approach: Synergy of sensors Multiple-Doppler radar analysis of 15 July 2007, additional validation with photos, lightning and MSG data Development of cloud-top height evaluated from radar and MSG data Analysis of microphysics of clouds by using polarimetric radar data > Folie 3 Folie 3

4 Initiation: Horizontal development parallax corrected positions of BT10.7µm 5 C ~ 80 km Radar sites multiple-doppler area > Folie 4 Folie 4

5 Life cycle and vertical development Decay Tropopause lapse rate: 0.6 K per 100 m Mature state discrepancy between radar cloud height and MSG related cloud height (ice shield) 3 km at 5 min -> vertical motion Initiation ~ 10 m/s > Folie 5 Folie 5

6 Life cycle and vertical development Tropopause lapse rate: 0.6 K per 100 m reference time steps for triple Doppler calculation 3 km at 5 min -> vertical motion ~ 10 m/s > Folie 6 Folie 6

7 Data radar volume scan time intervals Feldberg radar Türkheim radar Karlsruhe radar reference time for Multiple Doppler calculation 14:30 14:37 14:30 14:37 14:30 14:34 14:35 14:45 14:52 14:45 14:52 14:50 14:54 14:50 15:00 15:07 15:00 15:07 15:00 15:04 15:05 15:15 15:22 15:15 15:22 15:20 15:24 15:20 Validation case POLDIRAD RHI scan Feldberg radar Türkheim radar Karlsruhe radar reference time for Multiple Doppler calculation 14:43-14:45 14:45 14:52 14:45 14:52 14:30 14:34 14:43 > Folie 7 Folie 7

8 Method Multiple Doppler calculation radar data interpolation by using REORDER (by NCAR) GRID distance in all 3 dimensions: 500m (tested: 300m, 1km) Cressman weighting scheme Determination of storm motion (best overlap of reflectivity) multiple doppler calculation by using CEDRIC (by NCAR) calculation horizontal wind components u,v smoothing of wind field calculation of vertical wind by using variation integration procedure with boundary conditions > Folie 8 Folie 8

9 Mid cloud flow structure at 5 km (msl) connect Result outflow in km out inflow km Contour of reflectivity (10 dbz interval, starting from 10 dbz) Horizontal wind vectors relative to the storm motion updraft downdraft > Folie 9 Folie 9

10 Statistics: Vertical wind from triple Doppler 14:35 14:50 Mean vertical velocity Standard deviation of vertical velocity 15:05 15:20 average of updrafts (w>0) average of downdrafts (w<0) Updraft higher than downdraft at the early phase (14:35) Dominating wind direction changes from updraft to downdraft: decreasing of storm intensity Mean values and standard deviation decreases at top and bottom (influence of boundary condition) Decreasing of absolute vertical wind speed indicate decay state Low border limited to 1.7 km: radar beam cannot reach the ground (15:20) > Folie 10 Folie 10

11 Validation first photo from serie: timestamp: 14:43 UTC location: POLDIRAD > Folie 11 Folie 11

12 elevation [ ] Validation azimuth [ ] add scale: elevation, azimuth > Folie 12 Folie 12

13 elevation [ ] Validation azimuth [ ] verify: Black Forest silhouette from SRTM topography data > Folie 13 Folie 13

14 elevation [ ] Validation azimuth [ ] add all lightning locations (cloud and ground stokes) 14:34 14:45 UTC photo time: 14:43 UTC > Folie 14 Folie 14

15 Validation motion direction 0 dbz elevation [ ] 10 dbz 20 dbz 30 dbz 40 dbz azimuth [ ] Reflectivity composit from triple-doppler analysis at 14:35 UTC (7 minutes before the photo was taken) > Folie 15 Folie 15

16 Validation 2 m/s elevation [ ] 12 m/s 7 m/s 2 m/s azimuth [ ] maximum updraft in line of sight from multiple-doppler analysis > Folie 16 Folie 16

17 Validation elevation [ ] -3 m/s -3 m/s -3 m/s azimuth [ ] maximum downdraft in line of sight from multiple-doppler analysis > Folie 17 Folie 17

18 Consistency check for triple Doppler results rms v 3D = 1 n v ir Measured i vr 3D 2 measured radial velocity from each radar site derived radial component corresponding to radar site from 3D triple Doppler calculation Higher dynamics (e.g. more turbulence) in the early Mature state? > Folie 18 Folie 18

19 elevation [ ] Scan position of POLDIRAD RHI scans azimuth [ ] 14:43 14:45 UTC > Folie 19 Folie 19

20 Quality check by using POLDIRAD Measured and retrieved radial velocities vr low ρhv : high particle variation in type, shape and orientation high σ: broad range of radial velocity values correlation coefficient σ lightning retrieved vr retrieved vr 5 m/s higher than measured 5 m/s lower than measured main part of discrepancy corresponds with occurrences of turbulence denoted by low ρhv and high spectral width > Folie 20 Folie 20

21 Estimation of hydrometeor content Estimation of hydrometeors by only using of ZDR and LDR after Höller et al Comparison: Estimation of hail signal (HDR) by using ZDR and Z after Aydin et al mixing zone (water, ice) evidence of hail below melting zone: HDR (Z+ZDR) classification (LDR+ZDR) hail spike at reflectivity > Folie 21 Folie 21

22 Temporal evolution of hydrometeor content 14:43 14:53 15:03 results method 15:13 counting of radar range bins for each hydrometeor type per altitude level (1km) normalizing by dividing through total count summarizing 5 RHI scans per time interval fraction of hail is negligible Early phase dominated by ice particles (snow, graupel) Mean altitude level of ice particles decreases with time, but ice shield seen at photos and MSG data not detected by radar > Folie 22 Folie 22

23 Cloud-top height Contours: Triple-radar derived cloud-top heights ( 8km, 12km, 13km ) Points: Parallax corrected cloud-top positions (MSG IR-channel) with color coded BT > Folie 23 Folie 23

24 Cloud-top height Discrepancy of cloud-top location during the dissipation state > Folie 24 Folie 24

25 Triple-Doppler results Maximum reflectivity in line of sight Radar cross section reflectivity horizontal wind > Folie 25 Folie 25

26 Conclusion Documentation of full life cycle by using radar data (inclusive multiple Doppler approach) MSG and lightning data Most active phase between 14:35 and 14:45 cloud extended to tropopause raised updraft, high cloud dynamic (turbulence), lightning, hail Temporal evolution of hydrometeor profiles (consistent with simulations?) Discrepancy between radar and satellite & photo due to small ice particles Future work: comparison with model results > Folie 26 Folie 26

27 Outlook: MesoNH-model results from E. Richard :35 14:50 15:05 w [m/s] 15: Spatial resolution: 500 m altitude level: 5 km Small spatial (or temporal) discrepancies between model and observation Different role of southern cell Same inflow and outflow characteristics > Folie 27 Folie 27

28 Thank you for your attention! > Folie 28 Folie 28

29 Statistics: Horizontal wind from triple Doppler Mean horizontal wind speed with standard deviation Mean wind direction derived from mean u and v wind components wind shear decreases during life time profiles of mean wind speed ~ constant standard deviation of mean horizontal wind correlated with cloud dynamics > Folie 29 Folie 29